Created by: neozhangthe1

Error in https://github.com/PaddlePaddle/DeepSpeech/blob/edaed68f33b7fff2455c26b880340ba680621f5c/decoders/swig/ctc_beam_search_decoder.cpp#L137 when decoding using beam search.

[INFO 2018-01-24 14:29:35,342 model.py:244] begin to initialize the external scorer for decoding [INFO 2018-01-24 14:29:35,439 model.py:255] language model: is_character_based = 1, max_order = 5, dict_size = 0 [INFO 2018-01-24 14:29:35,439 model.py:256] end initializing scorer. Start decoding ...

* Aborted at 1516804178 (unix time) try "date -d @1516804178" if you are using GNU date *

PC: @ 0x0 (unknown)

* SIGFPE (@0x7fe13bdd2f36) received by PID 8361 (TID 0x7fe12114b700) from PID 1004351286; stack trace: *

@ 0x7fe16d7e4390 (unknown) @ 0x7fe13bdd2f36 PathTrie::iterate_to_vec() @ 0x7fe13bdd2ed8 PathTrie::iterate_to_vec() @ 0x7fe13bdc8992 ctc_beam_search_decoder() @ 0x7fe13bdc9bd7 std::_Function_handler<>::_M_invoke() @ 0x7fe13bdc9589 std::__future_base::_State_baseV2::_M_do_set() @ 0x7fe16d7e1a99 __pthread_once_slow @ 0x7fe13bdc98ef std::__future_base::_Task_state<>::_M_run() @ 0x7fe13bdcac11 _ZNSt6thread5_ImplISt12_Bind_simpleIFZN10ThreadPoolC4EmEUlvE_vEEE6_M_runEv @ 0x7fe14b51ac80 (unknown) @ 0x7fe16d7da6ba start_thread @ 0x7fe16d5103dd clone @ 0x0 (unknown)

Floating point exception (core dumped)

The model is trained on an internal mandarin speech dataset.

trainer_count = 1 beam_size = 500 num_proc_bsearch = 8 num_proc_data = 8 num_conv_layers = 2 num_rnn_layers = 3 rnn_layer_size = 2048 alpha = 1.0 beta = 2.329 cutoff_prob = 1.0 cutoff_top_n = 40 use_gru = True use_gpu = True share_rnn_weights = False lang_model_path = '/mnt/models/lm/zh_giga.no_cna_cmn.prune01244.klm' decoding_method = 'ctc_beam_search' error_rate_type = 'cer' specgram_type = 'linear'